Area Suitability Prediction for Conserving Elephants: An Application of Likelihood Ratio Prediction Model

• Published in: Tropical agricultural research Vol. 25; no. 3; pp. 345 - 357
• Main Authors: Marasinghe, M.S.L.R.P., Dayawansa, N.D.K., Silva, R. P. De
• Format: Journal Article
• Language: English
• Published: Postgraduate Institute of Agriculture, University of Peradeniya 21.10.2015
• Subjects: habitat preference moelling, human-elephant conflict, likelihood ratio, spatial modelling
• ISSN: 1016-1422
• DOI: 10.4038/tar.v25i3.8044

Asian elephant (Elephas maximus) is the largest terrestrial mega herbivore in Asia. Their distribution is highly fragmented and occurs only in thirteen countries in the world. Survival of the Asian elephant is endangered mainly due to the habitat loss. Implementation of development projects in areas where elephants are habitually present is a challenge and rational location of the development projects avoiding elephant habitats requires considerable amount of resources and time. Therefore, development of a cost effective methodology for evaluating the existing and potential elephant habitats is a necessity. Ecological factors affecting elephant behaviour are spatial in nature. Use of the likelihood ratio model with geo-informatics enables evaluation of the effects of spatial factors and their interrelations on occurrence of a certain event or a phenomenon. The objective of this study was to employ Geographic Information System (GIS) and Remote Sensing techniques in evaluating the suitability of a given area for elephants using likelihood ratio prediction model. The study was carried out in the north-western and Mahaweli wildlife region of the Department of Wildlife Conservation using two elephant herds. The percentage slope, Normalized Difference Vegetation Index (NDVI), Normalized Difference Water Index (NDWI) and Normalized Difference Built-up Index (NDBI) were used to reflect the factors such as terrain, food and water availability and manmade disturbances, which influence on the habitat preference of elephants. The predicted elephant preference area map was compared and accuracy was tested with the actual preference area map developed using the telemetry data on movement of the two herds. It was found that the likelihood ratio prediction model could be used in predicting area suitability of elephants with 79 % accuracy. As the prediction shows an acceptable level of accuracy, the model could be used in practical decision making and investment planning. This model could be further improved by including variables that represents manmade disturbances more effectively. Tropical Agricultural Research Vol. 25 (3): 345-357 (2014)